dlc / folderol Goto Github PK

Need to handle timeouts gracefully.

The current feeds() method simply returns a list of every feed, which is OK but not ideal, especially when there are a large number of feeds. Need to implement and intelligent method that returns a list of feeds that are likely to have been updated, based on update history of each feed and the last time it was updated.

getstore just fails if the directory to which it's supposed to write doesn't exist.

Create documentation:

• Configuration, including multiple sites in a single YAML file
• Overview of Template Toolkit
• Detailed description of all available template parameters

Copy static files (configurably) from the input directory, similar to Planet's bill_of_materials

Example:

<entry>
<title>Sign Failure</title>
<link rel="alternate" href="http://feedproxy.google.com/~r/shipmentoffail/~3/RMx9D5q_DiQ/" />
<id></id>
<updated></updated>
<issued></issued>

Right now, the parser blindly continues if the feed it's parsing is not valid XML, which causes the next step to explode. ("Can't call method foo on an undefined value")

Add a field to the db for expanded_link, which will contain the followed version of the link for an entry. When expanding a link, remove utm_ parameters.

Turn this on and off by a config parmeter, defaulting to off.

The following should work as expected:

template_map:
    foo.tt2: foo/index.html

Also:

template_map:
    daily.tt2: $YEAR/$MONTH/$DAY/index.html
    tags.tt2: $TAGS/index.html

Anything proceeded by $ will need to be expanded at run time (which allows forward and backward references to variables), and will need to be array-away; a variable defined as:

tags:
    - foo
    - bar

Will need to be automatically iterated over.

This will also require some predefined variables, like $YEAR, $MONTH, and $DAY.

Right now, there is no way to manage items in the database, other than direct manipulation of the reader.db file. Need a tool to manage the database, for example to purge older entries, remove feeds entirely (even if a feed is no longer being fetched, items remain in the db and so show up in the generated files), and so on.

Include several example templates, that generate common output types, including OPML and an Atom feed.

Non-200 responses should be ignored, and not written to the cache.